WORLD JOURNAL OF SURGICAL ONCOLOGY

Wang et al. World Journal of Surgical Oncology 2014, 12:365http://www.wjso.com/content/12/1/365

RESEARCH Open Access

Role of apparent diffusion coefficients withdiffusion-weighted magnetic resonance imagingin differentiating between benign and malignantbone tumorsTingting Wang1, Xiangru Wu2†, Yanfen Cui1, Caiting Chu1, Gang Ren1† and Wenhua Li1*

Abstract

Background: Benign and malignant bone tumors can present similar imaging features. This study aims to evaluatethe significance of apparent diffusion coefficients (ADC) in differentiating between benign and malignant bonetumors.

Methods: A total of 187 patients with 198 bone masses underwent diffusion-weighted (DW) magnetic resonance(MR) imaging. The ADC values in the solid components of the bone masses were assessed. Statistical differencesbetween the mean ADC values in the different tumor types were determined by Student’s t-test.

Results: Histological analysis showed that 84/198 (42.4%) of the bone masses were benign and 114/198 (57.6%)were malignant. There was a significant difference between the mean ADC values in the benign and malignantbone lesions (P <0.05). However, no significant difference was found in the mean ADC value between non-ossifyingfibromas, osteofibrous dysplasia, and malignant bone tumors. When an ADC cutoff value ≥1.10 × 10−3 mm2/swas applied, malignant bone lesions were excluded with a sensitivity of 89.7%, a specificity of 84.5%, a positivepredictive value of 82.6%, and a negative predictive value of 95.3%.

Conclusions: The combination of DW imaging with ADC quantification and T2-weighted signal characteristics ofthe solid components in lesions can facilitate differentiation between benign and malignant bone tumors.

Keywords: Apparent diffusion coefficient, Bone tumors, Differentiation, Diffusion-weighted imaging

BackgroundPreoperative characterization of benign, malignant, andtumor-like bone lesions is important in order to makeinformed choices regarding treatment strategies. Al-though plain film radiography is still considered thefirst-line imaging modality for assessing the nature anddefining the characteristics of primary bone lesions, cer-tain areas of the musculoskeletal system may be difficultto profile in plain films due to overlapping structures[1]. Computerized tomography can provide more de-tailed information, including focal destruction, periosteal

* Correspondence: wenhualimyj@sohu.com†Equal contributors1Department of Radiology, Xinhua Hospital affiliated to Shanghai Jiao TongUniversity School of Medicine, 1665 Kong Jiang Road, Shanghai 200092,ChinaFull list of author information is available at the end of the article

© 2014 Wang et al.; licensee BioMed Central LCommons Attribution License (http://creativecreproduction in any medium, provided the orDedication waiver (http://creativecommons.orunless otherwise stated.

reaction, subtle matrix mineralization, and endostealscalloping [2]. Magnetic resonance (MR) imaging is con-sidered the most advanced imaging technique and themost sensitive for evaluating changes in bone-marrowand defining the extent of a lesion, particularly whenplain films or computerized tomography findings aresuboptimal or indeterminate; however, it is not alwaysthe most specific. Despite their respective advantages,none of these imaging techniques can reliably differen-tiate between benign and malignant bone tumors, asmany lesions are non-specific and display varying im-aging characteristics on T1- and T2-weighted images.MR characterization of bone lesions can be improved bythe use of MR diffusion-weighted (DW) imaging, as thisis sensitive to changes in the microdiffusion of waterinto both intracellular and extracellular spaces. The

td. This is an Open Access article distributed under the terms of the Creativeommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andiginal work is properly credited. The Creative Commons Public Domaing/publicdomain/zero/1.0/) applies to the data made available in this article,

Wang et al. World Journal of Surgical Oncology 2014, 12:365 Page 2 of 6http://www.wjso.com/content/12/1/365

advantage of evaluating diffusion is the ability to probethe cellularity of neoplasms. Apparent diffusion co-efficients (ADC) are largely proportional to the ratio ofextracellular and intracellular components, cell density,intracellular organelles, matrix fibers, and soluble mac-romolecules. Tumors with different levels of cellularityhave different ADC values corresponding to changes inrestricted diffusion [3-6].The purposes of this study were to clarify the relation-

ship between ADC values in the solid components ofbone masses and to evaluate its supplementary use in dif-ferentiating between benign and malignant bone tumors.

MethodsPatient selectionThis study received approval from our institutional(Xinhua Hospital) review board and the requirement toobtain written informed consent was waived. A total of178 patients who had been diagnosed with bone tumorsbetween January 2005 and March 2014 were enrolled inthis study. The selection criteria were as follows: thediagnosis was confirmed by histological biopsy or surgery,MR imaging was performed using a 3.0 T magnet, andboth conventional MR imaging with DW imaging andcontrasted-enhancement MR imaging were performed. Aretrospective evaluation of the MR imaging data wasundertaken.

MR imaging protocolAll patients underwent MR imaging with a 3.0T MR unit(GE Medical Systems; Milwaukee, WI, USA). Axial non-contrast T1-weighted (TR/TE, 400 to 500/10 to 12 ms) andaxial T2-weighted (TR/TE, 4,000 to 5,000/100 to 120 ms)imaging were performed with chemical shift-selective fatsaturation pulse using the following parameters: slice thick-ness, 5 mm; gap, 1 mm; field of view, 20 to 40 cm; matrix,256 × 256; and excitation, 2. Sagittal T1-weighted and T2-weighted (TR/TE, 3,000 to 5,000/100 to 110 ms) fast spin-echo imaging without chemical shift-selective fat saturationpulse were also performed using the parameters describedabove. DW-MR imaging was performed in the axial or sa-gittal plane prior to administration of contrast mediumusing a single-shot echo-planar imaging sequence (TR/TEeffective range, 6,000 to 8,000/70 to 100 ms; slice thickness/intersection gap, 5/1 mm; field of view, 20 to 40 cm; matrix,128 × 128; excitation, 2. A b-value of 0 and 1,000 s/mm2

were also applied in three orthogonal directions. Post-contrast-enhanced axial and sagittal T1-weighted imagingwere also performed using the parameters described abovewith the exception of the 5 mm slice thickness.

MR and MR-DW image analysisConventional MR and DW-MR imaging data were ana-lyzed on an Advantage Windows workstation 4.2 (GE

Healthcare, Milwaukee, WI, USA). Image analysis wascarried out by two radiologists in consensus (with 7 and9 years’ experience in musculoskeletal MR imaging, re-spectively). The signal intensity of the solid componentson T2-weighted MR images was defined as intermediateor high, relative to the muscle signal and signal intensityof the solid portions exhibiting enhancement post-injection. The signal intensity of the solid components atb = 1,000 s/mm2 on the DW images was defined asintermediate or low relative to that of the muscle.

Data calculation and analysisThe solid components of the lesions were identified onT2-weighted and post-contrast T1-weighted images, andwere matched on ADC maps. The ADC values of thesolid components in each tumor were measured on DWimages by a radiologist using an Advantage Windowsworkstation 4.2 and FuncTool software (GE MedicalSystems). In order to minimize variability, the largestpossible region of interest (ROI) was placed manually inthe solid part of the tumor in each image (range: 10 to80 mm2). If the lesion exhibited irregular or heteroge-neous solid components, two or three ROIs were drawnwithin the targeted components and the mean ADCvalue was calculated for the analyses.

Statistical analysisAll analyses were performed using SPSS v. 13.0 softwarefor Windows (SPSS; Chicago, IL, USA). Differences in themean ADC values of the bone tumors between the benignand malignant groups were evaluated using Student’st-test. A value <0.05 was considered statistically sig-nificant. Biopsy or surgical pathology results were used asreference standards for assessment of the bone tumors.Receiver operating characteristic curve analysis was per-formed to assess the diagnostic performance of the meanADC values in characterization of benign and malignantbone tumors.

ResultsDemographics and histopathological characteristicsThe histopathological types of 198 bone masses in the178 patients are summarized in Table 1. The studygroup consisted of 81 males and 97 females, with amean age of 31.52 ± 28.31 years (range: 1 to 92 years).These included 131 patients with solitary bone tumors,22 patients with 28 Langerhans cell histiocytosis, and 25patients (n = 15 lung cancer, n = 4 breast cancer, n = 3prostate cancer, and n = 3 colorectal cancer) with 39metastatic bone tumors. A total of 84/198 (42.4%) bonemasses were benign and 114/198 (57.6%) were malig-nant. The diameters of the lesions in the benign groupwere 1.5 to 11.2 cm (median: 3.7 cm) and 1.2 to 16 cm(median, 4.1 cm) in the malignant group.

Table 1 Histological type and apparent diffusion coefficient (ADC) values of 198 bone masses (mean ± SD× 10−3 mm2/s)

Type of bone masses No. of lesions High SI on T2WI (%) High SI on DWI (%) Range of ADC values Mean ADC value

Benign bone lesions 84 54 (64.3) 50 (59.5) 0.49–1.59 1.17 ± 0.36

Non-ossifying fibroma 17 4 (23.5) 2 (11.8) 0.49–0.99 0.78 ± 0.17

Osteofibrous dysplasia 18 3 (16.7) 1 (5.6) 0.89–1.19 0.97 ± 0.17

Chondromyxoid fibroma 6 5 (83.3) 4 (66.7) 1.18–1.59 1.33 ± 0.15

Langerhans cell histiocytosis 28 28 (100) 28 (100) 0.96–1.55 1.29 ± 0.18

Giant cell tumor of bone 15 14 (93.3) 15 (100) 0.98–1.47 1.21 ± 0.20

Malignant bone lesions 114 101 (88.6) 105 (92.1) 0.58–1.35 0.86 ± 0.20

Chordoma 27 21 (77.8) 24 (88.9) 0.59–1.11 0.80 ± 0.14

Ewing sarcoma 9 6 (66.7) 7 (77.8) 0.70–0.91 0.82 ± 0.07

Osteosarcoma 8 6 (75.0) 5 (62.5) 0.87–1.02 0.97 ± 0.08

Chondrosarcoma 12 10 (83.0) 11 (91.7) 0.69–1.21 0.94 ± 0.15

Plasmacytoma 8 8 (100) 8 (100) 0.68–1.12 0.86 ± 0.15

Primary lymphoma 11 11 (100) 11 (100) 0.68–1.09 0.87 ± 0.14

Metastatic bone tumor 39 39 (100) 39 (100) 0.58–1.12 0.81 ± 0.14

SI, Signal intensity; T2WI, T2-weighted image; DWI, Diffusion-weighted image.

Wang et al. World Journal of Surgical Oncology 2014, 12:365 Page 3 of 6http://www.wjso.com/content/12/1/365

Relationship between DW imaging and bone tumor typesThe percentage of T2-weighted MR images of the solidcomponents in bone masses giving intermediate or highsignal intensities was significantly lower in the benigntumor group compared to the malignant tumor group(64.3% and 88.6%, respectively; P <0.01). The 84 benignlesions with homogeneous or heterogeneous low signalintensities included 76.5% non-ossifying fibromas, 83.3%osteofibrous dysplasia tumors, 16.7% chondromyxoidfibromas, and 6.7% giant cell tumors of bone.Evaluation of the bone masses by DW imaging at

b = 1,000 s/mm2, revealed that a significantly higher pro-portion of malignant lesions exhibited a high signal in-tensity within the solid components of the bone massescompared to those in benign lesions (92.1% and 59.5%, re-spectively; P <0.05). Further analysis of the images showedthat the presence of a solid component with high signalintensity on T2-weighted images and high signal intensityon DW images with low ADC values (<1.10 × 10−3 mm2/s,at b = 1,000 s/mm2) could be considered predictive of ma-lignancy (Figure 1). Conversely, the presence of a solidcomponent with high or low signal intensity on T2-weighted images and low signal intensity on DW imageswith high ADC values (≥1.10 × 10−3 mm2/s, at b = 1,000 s/mm2), or low signal intensity on T2-weighted images andlow signal intensity DW images with low ADC values(<1.10 × 10−3 mm2/s, at b = 1,000 s/mm2) could be con-sidered predictive of a benign mass (Figure 2).

ADC analysisThe mean ADC values of the solid components in thebone masses were determined for each group. There wasconsiderable overlap in the range of values observed within

the benign and malignant bone tumors; however, the meanADC value for benign tumors (1.17 ± 0.36 × 10−3 mm2/s)was significantly higher than that in malignant tumors(0.87 ± 0.20 × 10−3 mm2/s; P <0.05). When the tumor sub-types were compared, no significant difference in theADC values was found between the non-ossifying fibro-mas or osteofibrous dysplasia and malignant bone tumors(P >0.05). When an ADC cutoff value ≥1.10 × 10−3 mm2/swas applied, the benign and malignant bone tumors couldbe differentiated with a sensitivity of 89.7%, a specificity of84.5%, a positive predictive value of 82.6%, and a negativepredictive value of 95.3%, suggesting that this may be theoptimal cutoff value for the ADC.

DiscussionOur results demonstrated that the presence of a solidcomponent in bone lesions with high signal intensity onDW and T2-weighted images, combined with low ADCvalues can be used to distinguish between malignant andbenign bone lesions. These findings suggested that DWimaging with quantitative analysis of ADCs may improvethe diagnostic performance of bone MR imaging, andprovide clinically valuable information on the tumormicroenvironment.DW imaging has been extensively investigated for its

ability to characterize tissue in various lesions. It has beenshown to increase the accuracy in distinguishing benignfrom malignant masses, and discriminating between meta-static and benign lymphadenopathies. It has also provedsuccessful in the evaluation of cerebral ischemia and intra-cranial tumors [1,3,4]. Malignant tumors with a high nu-clear/cytoplasmic ratio, hypercellularity, and a reducedextracellular matrix often have restricted mobility of water

Figure 1 A 7-year-old girl with an eosinophilic granuloma. (A) Axial T1-weighted image showing an isointense small round mass within theoccipital bone (arrow). (B) An axial T2-weighted image reveals that the mass is hyperintense (arrow). (C) Axial DW imaging reveals that the massis slightly hyperintense with a high ADC value (circled area: ADC = 1.50 × 10−3 mm2/s) (arrow). (D) Axial enhanced T1-weighted image showingthe lesion with marked enhancement (arrow).

Wang et al. World Journal of Surgical Oncology 2014, 12:365 Page 4 of 6http://www.wjso.com/content/12/1/365

molecules and low ADC values, whereas benign tumorsgenerally have higher ADC values. These variations werereflected in our results. Although there was some overlapin the ADC values between the malignant and benignbone lesions, the mean ADC values in the 114 malignantbone tumors was significantly lower than that in the 84benign bone masses. Our results further revealed that theoptimal ADC cutoff value for differentiating between be-nign and malignant tumors was 1.10 × 10−3 mm2/s. Thisresult was consistent with those given in previous reports[7-12].A study by Hayashida et al. [4] evaluated the contribu-

tion of DW imaging in combination with quantitative ana-lysis of ADCs in the characterization of 20 bone masses,including 8 solitary bone cysts, 5 fibrous dysplasia tumors,and 7 chondrosarcomas. Their results suggested that thismethod of imaging bone lesions was not suitable fordifferentiating between benign and malignant bone le-sions. This apparent discrepancy may have been due to

differences in pathological bone architecture. However,our results revealed a significant difference in the meanADC values of the solid components between malignantand benign bone lesions. In addition, our results showedlow ADC values in 17 non-ossifying fibromas and 18osteofibrous dysplasia tumors. These may have been dueto the presence of abundant collagen-producing fibroblas-tic cells and a dense network of collagen fibers within theextracellular matrix, which can restrict the Brownian mo-tion of water molecules. This characteristic may also havebeen responsible for the absence of a significant differencebetween the mean ADC values of non-ossifying fibromasor osteofibrous dysplasia tumors and malignant bonelesions.Our results showed that 88.6% of the T2-weighted im-

ages of solid components in malignant tumors displayedhigh signal intensities, compared to 11.4% which dis-played low signal intensities. However, this value in-creased to 92.1% when the DW images with low ADC

Figure 2 A 23-year-old man with an osteosarcoma. (A) An axial T1-weighted image showing a hypointense left tibial osteosarcoma mass(arrow). (B) Axial T2-weighted image showing that the mass has high signal intensity (arrow). (C) Axial DW imaging reveals that the mass ishyperintense with a low ADC value (circled area: ADC = 0.985 × 10−3 mm2/s). (D) Axial contrast-enhanced T1-weighted image shows markedenhancement of the bone mass.

Wang et al. World Journal of Surgical Oncology 2014, 12:365 Page 5 of 6http://www.wjso.com/content/12/1/365

values were analyzed. Bone lesions which exhibited lowsignal intensity on T2-weighted images and high signalintensity on DW images may have resulted from solidcomponents with increased nuclear/cytoplasmic ratiosand hypercellularity due to a reduction in both theextracellular matrix and the diffusion space of water pro-tons in the extracellular and intracellular dimensions. Inaddition, the low signal intensity on T2-weighted imagesand DW images of benign bone lesions, such as non-ossifying fibromas and osteofibrous dysplasia tumors,may have been due to the high density of fibers, low cel-lularity, and low water content in both the extracellularand intracellular spaces [4,13]. Our findings confirmedthat a high signal intensity on DW images of solid com-ponents with low ADC values can serve as a useful cri-terion for predicting malignancy in bone lesions, andthat a low signal intensity on T2-weighted images andDW images of solid components with low ADC valuesmay be an effective criterion for predicting the presenceof benign disease.Our study had the following limitations: DW imaging

often has poor spatial resolution, therefore, drawing theROI on DW images while viewing T2-weighted or con-trasted T1-weighted images may result in informationbias. A better approach may be to fuse the DW imagesof solid components that show an abnormal signal atb = 1,000 × 10−3 mm2/s onto structural images in orderto accurately position the ROI. Furthermore, the proposedvalue for the ADC threshold will need to be validated in alarger group of patients, as the ADC can be affected bymany factors, including magnetic susceptibility, spatial

resolution, signal to noise ratio, and the pathophysio-logical characteristics of the bone lesions.

ConclusionsDW imaging is a potentially valuable method for differen-tiating between benign and malignant bone tumors as itoffers both high sensitivity and specificity. Bone masseswith solid components that exhibit low signal intensity onT2-weighted and DW images and low ADC values are in-variably benign.

AbbreviationsADC: Apparent diffusion coefficients; DW: Diffusion-weighted; MR: Magneticresonance; ROI: Region of interest.

Competing interestsThis research received no specific grant from any funding agency in thepublic, commercial, or not-for-profit sectors. The authors declare that theyhave no conflicts of interest.

Authors’ contributionsWH Li planned, designed, and analyzed the study. TT Wang collected dataand wrote the manuscript. YF Cui and G Ren collected data. CT Chudesigned the study. XR Wu contributed as a pathologist and helped in thewriting of the manuscript. All authors read and approved the finalmanuscript.

AcknowledgementsWe thank Huarong Gong, Huitong AN, Qiufeng Yin, and Ming Liu for theirtechnical support.

Author details1Department of Radiology, Xinhua Hospital affiliated to Shanghai Jiao TongUniversity School of Medicine, 1665 Kong Jiang Road, Shanghai 200092,China. 2Department of Pathology, Xinhua Hospital affiliated to Shanghai JiaoTong University School of Medicine, 1665 Kong Jiang Road, Shanghai200092, China.

Wang et al. World Journal of Surgical Oncology 2014, 12:365 Page 6 of 6http://www.wjso.com/content/12/1/365

Received: 23 April 2014 Accepted: 18 November 2014Published: 29 November 2014

References1. Wyers MR: Evaluation of pediatric bone lesions. Pediatr Radiol 2010,

40:468–473.2. Woertler K: Benign bone tumors and tumor-like lesions: value of

cross-sectional imaging. Eur Radiol 2003, 13:1820–1835.3. Pearce T, Philip S, Brown J, Koh DM, Burn PR: Bone metastases from

prostate, breast and multiple myeloma: differences in lesion conspicuityat short-tau inversion recovery and diffusion-weighted MRI. Br J Radiol2012, 85:1102–1106.

4. Hayashida Y, Hirai T, Yakushiji T, Katahira K, Shimomura O, Imuta M, NakauraT, Utsunomiya D, Awai K, Yamashita Y: Evaluation of diffusion-weightedimaging for the differential diagnosis of poorly contrast-enhanced. J MagReson Imaging 2006, 23:377–392.

5. Moon WJ, Lee MH, Chung EC: Diffusion-weighted imaging with SensitivityEncoding (SENSE) for detecting cranial bone marrow metastases:comparison with T1-weighted images. Korean J Radiol 2007, 8:185–191.

6. Jaramilo D: Whole-body MR, imaging, bone diffusion imaging: how andwhy? Pediatr Radiol 2010, 40:978–984.

7. Ginat DT, Mangla R, Yeaney G, Johnson M, Ekholm S: Diffusion-weightedimaging for differentiating benign from malignant skull lesions andcorrelation with cell density. Am J Roentgenol 2012, 198(6):597–601.

8. Budde MD, Gold E, Jordan EK, Frank JA: Differential microstructure andphysiology of brain and bone metastases in a rat breast cancer modelby diffusion and dynamic contrast enhanced MRI. Clin Exp Metastasis2012, 29(1):51–62.

9. Garrett KM, Kim HK, Stanek J, Emery KH: MR findings of primary bonelymphoma in a 15-year-old girl: emphasis on diffusion-weightedimaging. Pediatr Radiol 2011, 41(5):658–662.

10. Hillengass J, Stieltjes B, Bäuerle T, McClanahan F, Heiss C, Hielscher T,Wagner-Gund B, Habetler V, Goldschmidt H, Schlemmer HP, Delorme S,Zechmann CM: Dynamic contrast-enhanced magnetic resonance imaging(DCE-MRI) and diffusion-weighted imaging of bone marrow in healthyindividuals. Acta Radiol 2011, 52(3):324–330.

11. Costa FM, Ferreira EC, Vianna EM: Diffusion-weighted magnetic resonanceimaging for the evaluation of musculoskeletal tumors. Magn ResonImaging Clin N Am 2011, 19(1):159–180.

12. Dremmen MH, Hofman PA, Hof JR, Stokroos RJ, Postma AA: The diagnosticaccuracy of non-echo-planar diffusion-weighted imaging in the detectionof residual and/or recurrent cholesteatoma of the temporal bone. AJNR AmJ Neuroradiol 2012, 33(3):439–444.

13. Ozgen B, Oguz KK, Cila A: Diffusion MR imaging features of skull baseosteomyelitis compared with skull base malignancy. AJNR Am JNeuroradiol 2011, 32(1):179–184.

doi:10.1186/1477-7819-12-365Cite this article as: Wang et al.: Role of apparent diffusion coefficientswith diffusion-weighted magnetic resonance imaging in differentiatingbetween benign and malignant bone tumors. World Journal of SurgicalOncology 2014 12:365.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit